Skip links

Automotive training and Validation data

DataSets

Cognata’s sensor modeling technology is tailored to Synthetic Datasets for sensor characteristics.

Different environments

Various weather

Traffic Light Datasets

Diverse, highly realistic, and fully configurable traffic light datasets that enrich real world training data and
improve traffic signal recognition in ADAS and autonomous vehicle applications.

Diverse, highly realistic, and fully configurable traffic light datasets that enrich real world training data and improve traffic signal recognition in ADAS and autonomous vehicle applications.

Diverse Catalog of Signals
Before
After
  • Our traffic signal housings consider every detail. We provide variable orientation, bulb count, tilt angle, visor and backplate configuration, size, and color for enriching any perception dataset.
Accurate Illumination
Before
After
  • Traffic light illuminators feature accurate color and luminance. They include global symbols for both left- and right-hand traffic configurations to augment and correct bias in any training data.
Customized Combinations
Before
After
  • Fully customized signal sets are simple. Diverse signal poles can be combined with any number of traffic lights and signs, in any configuration for your specific dataset needs.

Pedestrian

Cognata synthetic data provides the quality enrichment needed to thoroughly train modern ADAS
and Autonomous Vehicle systems for vulnerable road user (VRU) detection applications.

Cognata synthetic data provides the quality enrichment needed to thoroughly train modern ADAS and Autonomous Vehicle systems for vulnerable road user (VRU) detection applications.

Diverse Pedestrians and Cyclists from around the World
  • Whether cyclists, jaywalking pedestrians, a wheelchair user, or a cycle rickshaw – vehicles share the road with countless vulnerable road users.
  • Cognata synthetic training data provides the diverse catalogs of pedestrians, bicyclists, and personal mobility needed for global coverage.
Roadworks and Emergency Response Scenarios
  • Work zone crashes kill an average of two people every day in the US alone, with those directing traffic at highest risk. Our datasets provide construction workers, police, and emergency first responders for safe robust virtual training of pedestrian detection for these safety-critical scenarios.
Configurable Pose and Placement
  • Because pedestrians and cyclists both move and change pose, they present an additional challenge for detection and classification systems. Cognata makes it easy to parameterize and automatically generate training data comprising vulnerable road users in a wide range of poses and positions within the scene.

Synthetic Vehicles

Provides a range of vehicle catalogs ranging from automobiles and two-wheeled vehicles
to trucks, buses, construction and emergency vehicles.

Provides a range of vehicle catalogs ranging from automobiles and two-wheeled vehicles to trucks, buses, construction and emergency vehicles.

Passenger Vehicle
  • Cars, motorcycles, parking workers, highways, You can create any Data Set, even a driver.
Trucks, Trailers and Buses
  • Pick-up trucks, buses, trailers, and more.
    Provides size and shape.
Emergency and Special Purpose
  • Construction, industrial and agricultural sectors, etc.
    Special purpose vehicles can also be implemented.